-

Inspiration

While several datasets for autonomous navigation have become available in recent years, they have focused on structured driving environments. This usually corresponds to well-delineated infrastructures such as lanes, a small number of well-defined categories for traffic participants, low variation in the object or background appearance, and firm adherence to traffic rules.

The state-of-the-art methods which have worked well on such structured environment datasets like Cityscapes do not tend to perform well on the recently compiled Indian Driving Dataset which is based on roads of Hyderabad and Bengaluru. In this project we try to perform semantic segmentation on this dataset in an attempt to solve the autonomous driving challenges on Indian roads.

What it does

Planned Expectations

The main goal is to build an efficient deep learning model to perform image segmentation for the data. To obtain level-1 labels as in label hierarchy which consists of mainly 7 class labels. The task is to predict pixel-wise mask for each object in the image that is the pixel-level prediction for a given image.

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for -

Built With

• cnn
• colab
• jupyter
• keras
• python
• resnet
• tensorflow
• unet